Session S53.4

Model-Based Estimation of Intracranial Pressure and Cerebrovascular Autoregulation

FM Kashif*, T Heldt, GC Verghese

Massachusetts Institute of Technology
Cambridge, MA, USA

Inadequate oxygen delivery to the brain -- even over short periods -- can impair or even irreversibly damage cerebral function. To buffer this extreme sensitivity to mismatches in oxygen demand and oxygen delivery, blood flow to the brain is normally tightly regulated through global and local cardiovascular control mechanisms. Failure of the cerebral vasculature to regulate blood flow is commonly seen in pathological conditions such as stroke and traumatic brain injury, and correlates strongly with disease severity. Current methodologies for monitoring cerebrovascular state rely on measurements of intracranial pressure (ICP), and are therefore reserved for only the sickest of patients due to their high level of invasiveness. In this study, we employ a mathematical model-based approach to estimate cerebrovascular properties and ICP as a function of time from signals that can be acquired entirely non-invasively.
In a first step, we adapt a well-established lumped-parameter mathematical model of the cerebrovascular system to allow for generation of pulsatile intracranial pressure waveforms, and waveforms of blood pressure and flow in various segments of the brain. The model captures autoregulatory mechanisms through flow-dependent vascular compliances and pressure-dependent vascular resistances. As input to the model, we supply experimental recordings of arterial blood pressure. We introduce transients in ICP, so-called 'plateau-waves', through expansion or reduction in cerebrospinal fluid volume.
In a second step, we employ a reduced-order model to estimate key cerebrovascular parameters and ICP, making use of time-scale separation of important cardiovascular dynamics.
Over a wide range of simulated pressure changes, our estimates track dynamic changes in ICP usually to within 10% of the reference values. Estimates of cerebrovascular parameters tend to fall within 5% to 10% of the references values. These results suggest that we can dynamically track the state of cerebral autoregulation.

(Abstract Control Number: 151)